The patch has been updated, the attached files will replace my previous
post completely against ruleutils.c 1.143.
pg_get_indexdef is extended to take 3 arguments:
pg_get_indexdef(index_oid, columnNr_int, prettyOpt_int)
Still, pg_get_indexdef(oid) will deliver the same result as it used to be.
Th
Is there a problem about the archive?
I posted this addition to the patch on July 2, and received it over the
patches mailing list. Still, it doesn't appear in cvs, in "unapplied
patches" or in the mailing list archive! So although that message made
its way through the list server, it didn't end
Hi Bruce,
so here's the complete patch against the current cvs.
Description:
The attached patches will add
pg_get_ruledef(oid, bool)
pg_get_viewdef(text, bool)
pg_get_viewdef(oid, bool)
pg_get_indexdef(oid, int4, bool)
pg_get_constraintdef(oid, bool)
pg_get_expr(text, oid, bool)
If the
Bruce Momjian wrote:
Andreas, looks good, but I need a diff -c, context diff.
Hi Bruce,
I intentionally only attached only non-context diffs because the patch
is about 50 % size of the original file. Now, here's the same as context
diff.
Regards,
Andreas
Index: pg_proc.h
===
Bruce Momjian wrote:
Tom Lane wrote:
Bruce Momjian <[EMAIL PROTECTED]> writes:
Tom, how do I pass PG_FUNCTION_ARGS to another function, while adding a
new parameter?
I wouldn't. Do the PG_GETARGS in the wrapper, and have the called
function take a normal C parameter list.
So I
Tom Lane wrote:
Andreas Pflug <[EMAIL PROTECTED]> writes:
+ int prettyFlags = !PG_ARGISNULL(1) && PG_GETARG_BOOL(1) ? PRETTYFLAG_PAREN|PRETTYFLAG_INDENT : 0;
Since the pg_proc entries are all marked strict, it's unnecessary for
you to write any ARGISNULL
Tom Lane wrote:
Applied with some editorializing. In particular, I don't believe the
original did the right thing with (a - (b - c)).
Oops, missed that case...
But now, we have (a + ( b + c)) again.
A patch that removes parentheses for + and * is appended.
Regards,
Andfdsa
Now the patch is *really* appended :-)
Tom Lane wrote:
Applied with some editorializing. In particular, I don't believe the
original did the right thing with (a - (b - c)).
Oops, missed that case...
But now, we have (a + ( b + c)) again.
A patch that removes parentheses for + and * is appende
Tom Lane wrote:
Now the patch is *really* appended :-)
And rejected.
Ok, the ckeck for node being the first child already does the trick for
standard l-t-r evaluation.
You cannot assume that an operator is commutative or
associative just because it has a name you think ought to be.
(For
Tom Lane wrote:
Andreas Pflug <[EMAIL PROTECTED]> writes:
Well, to me it's not well-known that floating-point addition is not
associative, do I need to re-learn my math?
regression=# select (1.0::float8 + (-1.0::float8)) + 1.0e-20::float8;
?column?
--
1e-20
(1 row)
The attached patch allows libpq to be compiled under native win32. It
adds the missing thread.c.
Additionally, it corrects the option to compile for debugging.
Regards,
Andreas
RCS file: /projects/cvsroot/pgsql-server/src/include/pg_config.h.win32,v
retrieving revision 1.11
diff -u -r1.11 pg_conf
The attached patch reenables non-blocking operation under win32.
The global variable was left uninitialized, so the compiler would
silently make it 0. This resulted in connectMakeNonblocking() setting
nonblocking to zero, i.e. blocking.
Regards,
Andreas
RCS file: /projects/cvsroot/pgsql-server
The attached patch enables libpq to be linked with ssl support (openssl
0.9.7.b tested).
Client certificates are commented out because the implementation is *nix
specific, regarding the location resolution of the .pem files. It needs
to be discussed where these files should be located. For NT/W2
The attached patch reenables non-blocking operation under win32.
The global variable was left uninitialized, so the compiler would
silently make it 0. This resulted in connectMakeNonblocking() setting
nonblocking to zero, i.e. blocking.
Regards,
Andreas
RCS file: /projects/cvsroot/pgsql-server/
Hm,
I don't see any reaction on my posts from last sunday, neither in this
list nor in unapplied patches, did they get lost?
If so, attached is a combined patch for all three problems with libpq
compiling under win32.
Regards,
Andreas
cvs diff -u interfaces\libpq\fe-connect.c interfaces\libpq\f
This was discussed in [HACKERS] TCP/IP with 7.4 beta2 broken?
I created a patch to hba.c which uses IPV4 entries as IPV6 entries if
running on a IPV6 system (which is detected from a port coming in as
AF_INET6).
192.168.0.0/24 -> :::102.168.0/120
192.168.0.0 255.255.255.0 ->
Kurt Roeckx wrote:
You're assuming all systems have an AF_INET6 constant, which is
not the case. Please make use of HAVE_IPV6.
Can't directly see anything else wrong with it.
Here's the patch with HAVE_IPV6 conditional compiling.
Regards,
Andreas
Index: hba.c
===
Hi Bruce,
+
+/* getpwuid doesn't exist under win32 */
+#define getpwuid(uid) NULL
+
#endif /* pg_config_h_win32__ */
Why was this needed? I realize we don't have getpwuid() on Win32, but
we do have GetUserName() for cases where we need the name but not the
directory.
Because all the getpwui
Andrew Dunstan wrote:
Andreas,
You should check that the CIDR mask is a valid integer. You would need
to use strtol() rather than atoi() to do that. Perhaps this should be
hoisted out of ip.c:SockAddr_cidr_mask() and put in hba.c.
Right, I added this.
Regards,
Andreas
Index: hba.c
===
Tom Lane wrote:
I thought this was still really messy, so I modified it to use a
separate "promote v4 address to v6" subroutine. I've applied the
attached patch (plus docs). It's not very well tested since I don't
have an IPv6 setup here; please check that it does what you want.
It SEGVs. Reason
Tom Lane wrote:
IMHO the struct needs to be created officially by
getaddrinfo(), which will lead more or less the the same solution I
posted previously.
No, we just need to use datastructures that are under our control for
the values that will get passed to rangeSockAddr. A few more lin
Larry Rosenman wrote:
Here's a quickie patch I did to fix it.
Patching this or not, if the function's called the backend will SEGV
either (at least on my 2.4.20 it does) because a IPV6 address is copied
in the memory space of a IPV4 address.
Regards,
Andreas
---(end o
Tom Lane wrote:
Andreas Pflug <[EMAIL PROTECTED]> writes:
Larry Rosenman wrote:
Here's a quickie patch I did to fix it.
Patching this or not, if the function's called the backend will SEGV
either (at least on my 2.4.20 it does) because a IPV6 address is copie
I wonder if it wouldn't be easier to reorder the TupDesc->attrs[] array
according to an attphysid when filling the TupDesc structure, right
after a column was dropped/recreated (before any indexes/constraints are
recreated), so attnum remains, while storage changes.
Example:
before:
attnum attph
My Deja wrote:
I am trying to get some query trees to appear in the PostgreSQL log
and in order to that I have set
client_min_messages = DEBUG1 in order to use the following settings
debug_print_parse, debug_print_rewritten, or debug_print_plan which
are required for the query tree to show up
When checking for thread safety with src/tools/thread/thread_test.c, the
mktemp function wants an argument that contains 6 X, while the current
version only supplies 5 X which will fail on my SuSE 8.1.
Patch attached.
Regards,
Andreas
Index: thread_test.c
Some modifications are needed to compile libpq with vc6:
fe-connect.c:
- pg_config_paths.h isn't available. SYSCONFDIR is already defined so
fe-connect.c doesn't need to include that.
patch appended
win32.mak:
- pg_config.h must be generated
- port/pgstrcasecmp.c is needed
- port/path.c is not ne
init_ssl_system will return 0 on success and -1 on failure, which will
be interpreted just the other way round in initialize_SSL.
Patch appended.
Regards,
Andreas
Index: fe-secure.c
===
RCS file: /projects/cvsroot/pgsql-server/src/i
Manfred Spraul wrote:
Btw, --enable-thread-safety on Linux (RedHat Fedora Core 1) fails in
configure with
configure: error:
*** Thread test program failed. Your platform is not thread-safe.
*** Check the file 'config.log'for the exact reason.
I had this too, for two reasons:
- configure checks
Tom Lane wrote:
Andreas Pflug <[EMAIL PROTECTED]> writes:
When checking for thread safety with src/tools/thread/thread_test.c, the
mktemp function wants an argument that contains 6 X, while the current
version only supplies 5 X which will fail on my SuSE 8.1.
Patch attached.
Isn&
Tom Lane wrote:
Andreas Pflug <[EMAIL PROTECTED]> writes:
init_ssl_system will return 0 on success and -1 on failure, which will
be interpreted just the other way round in initialize_SSL.
Patch appended.
Hmm, that looks backwards to me too, but this would seem to imply that
M
Bruce Momjian wrote:
fe-connect.c:
- pg_config_paths.h isn't available. SYSCONFDIR is already defined so
fe-connect.c doesn't need to include that.
patch appended
This shouldn't be needed anymore. Where is SYSCONFDIR coming from? Is
it from some Win32 include file? It should only be c
Bruce Momjian wrote:
Andreas Pflug wrote:
Bruce Momjian wrote:
fe-connect.c:
- pg_config_paths.h isn't available. SYSCONFDIR is already defined so
fe-connect.c doesn't need to include that.
patch appended
This shouldn't be needed anymore. Wher
Andreas Pflug wrote:
The attached patch will create a dummy pg_config_paths.h.
Additionally, ENABLE_THREAD_SAFETY is supported by the makefile (but
not by the sources, which need some rework)
Now including the patch...
Regards,
Andreas
Index: win32.mak
The appended patch implements ENABLE_THREAD_SAFETY for win32 compiled
with Visual C. Two additional files will supply the needed pthread stuff.
There's no SIGPIPE for native win32, so there are some #IFNDEF WIN32 to
skip installing a signal handler for this. I'm not sure if this is going
to sho
While trying to
pg_constraint JOIN pg_index ON indrelid=conrelid AND conkey=indkey
I noticed (once again) that these columns have different types (although
describing the same thing), and there's no equality operator for them.
The attached patch adds an equality operator
bool int2array_int2vec
Tom Lane wrote:
Andreas Pflug <[EMAIL PROTECTED]> writes:
The attached patch adds an equality operator
bool int2array_int2vector_eq(int2[], int2vector)
This seems like a remarkably specialized kluge ... gotta be a
better way ...
How? This could be coded to support any elemen
Magnus Hagander wrote:
Specifically about the logs, I still think there is a lot of value to
being able to read the logs remotely even if you can't restart
postmaster.
Since I believe that retrieving the logs easily without server file
access is a feature that's welcomed by many users, here's my p
Tom Lane wrote:
Bruce Momjian <[EMAIL PROTECTED]> writes:
Looks good to me. The only issue I saw was that the default file name
mentioned in postgresql.conf doesn't match the actual default.
I'm really not happy with the concept that the postmaster overrides
its stderr direction.
I agre
Sorry I didn't get back on this earlier, yesterday morning my internet
access was literally struck by lightning, I'm running temporary hardware
now.
Bruce Momjian wrote:
Looks good to me. The only issue I saw was that the default file name
mentioned in postgresql.conf doesn't match the actual d
Bruce Momjian wrote:
I have looked over this patch. I noticed this:
-static pthread_mutex_t init_mutex = PTHREAD_MUTEX_INITIALIZER;
-
+static pthread_mutex_t init_mutex;
+static int mutex_initialized = 0;
+if (!mutex_initiali
Bruce Momjian wrote:
Agreed. My pthread book says pthread_mutex_init() should be called only
once, and we have to guarantee that. If the Windows implentation allows
it to be called multiple times, just create a function to be called only
by Win32 that does that and leave the Unix safe.
Ok, so
Bruce Momjian wrote:
I was thinking of close/reopen so log files
could be rotated.
Log file rotation is fine, if we find a consensus quite soon how to
implement it... Seems as if I might find some time to implement it until
feature freeze.
The attached patch has the default filename issue fix
Tom Lane wrote:
This has got portability issues (fopen("ab"))
My doc says b is ignored on ansi systems, and recommends using it. Do
you have other experiences?
and I don't care for its
use of malloc in preference to palloc either.
Do we already have an applicable memory context in the postmast
Manfred Spraul wrote:
Wrong. There are portable test-and-set functions in the Win32 SDK: for
example InterlockedCompareExchange. They operate on ints and do not
need initialization.
'k, missed that one. Lets use it.
There is a third option: Add one C++ file and use the constructor to
initialize
So here's the updated ENABLE_THREAD_SAFETY patch for win32, to be
compiled under VC5/6/7 (tested with VC6).
nmake -f win32.mak [DEBUG=1] [USE_SSL=1] [ENABLE_THREAD_SAFETY=1]
are supported.
Regards,
Andreas
/*-
*
* pthread-win
Dave Page wrote:
It could still be run on NT4 under the following conditions:
1) Running as a service
2) Running if the user logged in is not an administrator.
Well, isn't "running as a service" sufficient? I thought
that was the only interesting case for non-hackers anyway.
As long as y
Dave Page wrote:
you can only run one
instance as a service on a single machine.
If you mean only run one instance of postmaster as service, that's not true.
If you like two pgsql servers (i.e. db clusters), you can install two
services, both using the same binary with different cmd line argume
Gavin Sherry wrote:
On Fri, 18 Jun 2004, Bruce Momjian wrote:
[snip]
TODO. You sound like a man who's expecting a
several-generations-polished facility when we only just committed
the first version today. I do not feel a need to have any of these
features in 7.5 ...
I just need to know
Gavin Sherry wrote:
On Fri, 18 Jun 2004, Andreas Pflug wrote:
Gavin Sherry wrote:
On Fri, 18 Jun 2004, Bruce Momjian wrote:
[snip]
TODO. You sound like a man who's expecting a
several-generations-polished facility when we only just committed
the first version today. I do not
Gavin Sherry wrote:
I would debate that.
Firstly, tablespaces aren't supported on windows yet.
Just a matter of time. And I'm talking of win32 workstations connecting
to *ix servers too.
Secondly, I'd think
that Unix users would be fine with a command line tool, especially one
that can connect to
Tom Lane wrote:
Andreas Pflug <[EMAIL PROTECTED]> writes:
IMHO checking objects in a tablespace is a routine administrative task,
so it should be supported natively by the server without need of
contribs.
I strongly disagree. Dropping a tablespace is not a routine activity,
Dr
Tom Lane wrote:
Andreas Pflug <[EMAIL PROTECTED]> writes:
Tom Lane wrote:
As for the authentication-is-expensive issue, what of it? You *should*
have to authenticate yourself in order to look inside another person's
database. The sort of cross-database inspection being pr
Dave Page wrote:
-Original Message-
From: [EMAIL PROTECTED] on behalf of Bruce Momjian
Sent: Sat 6/19/2004 1:05 AM
To: Andreas Pflug
Cc: Tom Lane; Gavin Sherry; PostgreSQL-patches
Subject: Re: [PATCHES] Tablespace patch review
We can build a gui on top of the command-line tool, no?
No
Bruce Momjian wrote:
I don't see why an admin tool can't connect to each database and get a
listing of what is in each tablespace. I don't think connecting to 100
databases to get that information will be slow.
Well, whatever you call slow or not slow.
I checked it; connecting 10 databases, re
Dave Page wrote:
-Original Message-
From: Andreas Pflug [mailto:[EMAIL PROTECTED]
Sent: Sat 6/19/2004 6:40 PM
To: Bruce Momjian
Cc: Dave Page; Tom Lane; PostgreSQL-patches
Subject: Re: [PATCHES] Tablespace patch review
Well, whatever you call slow or not slow.
I checked it; connecting
Two minor tweaks:
in libpq-be.h isn't present on win32 vc6. Apparently, it
isn't used in Linux either; libpq will compile without it. All usages of
gettimeofday throughout the backend don't seem connection related, so
libpq-be.h seems an odd place to include it.
ERROR in elog.h has conflicts w
Tom wrote:
It's there to declare struct timeval, and I'm fairly certain that diking
it out of the header would break things on some platforms. Where does
Windows define struct timeval?
struct timeval is defined in winsock.h under vc6.
I'm checking for _MSC_VER now.
Agreed. We define FRONTEND
The attached patch includes serverlog rotation with minimal shared
memory usage as discussed and functions to access it.
Regards,
Andreas
Index: doc/src/sgml/func.sgml
===
RCS file: /projects/cvsroot/pgsql-server/doc/src/sgml/func.sg
From an idea of Bruce, the attached patch implements the function
pg_tablespace_databases(oid) RETURNS SETOF oid
which delivers as set of database oids having objects in the selected
tablespace, enabling an admin to examine only the databases affecting
the tablespace for objects instead of scanni
Andreas Pflug wrote:
From an idea of Bruce, the attached patch implements the function
pg_tablespace_databases(oid) RETURNS SETOF oid
which delivers as set of database oids having objects in the selected
tablespace, enabling an admin to examine only the databases affecting
the tablespace for
Fabien COELHO wrote:
Dear patchers,
I have two minor patches that are being submitted but which do not appear
yet in the "official" patch queue on the web site:
http://momjian.postgresql.org/cgi-bin/pgpatches
That site is maintained by Bruce, who is out to Armenia until next week,
with virtua
Joe Conway wrote:
If there are no objections, I'll review and apply this tonight (west
coast USA time).
Andreas, please provide a corresponding documentation patch.
Here we are.
Since I don't have a SGML build environment, syntax is not checked.
Regards,
Andreas
Index: func.sgml
=
Joe Conway wrote:
Attached is the patch I plan to apply. There are a couple of changes
from what was posted.
1) You must have meant tablespace instead of namespace here:
+
+
pg_tablespace_databases(namespace_oid)
+ setof
Tom Lane wrote:
Andreas Pflug <[EMAIL PROTECTED]> writes:
The attached patch includes serverlog rotation with minimal shared
memory usage as discussed and functions to access it.
This patch is still unsafe and unworkable. Why would you make the
mechanism dependent on shared memor
Updated version.
Only timestamp of fresh logfile in shared mem, with sanity checks.
On SIGHUP, timestamp is checked if rotation was issued, as well as
changed log_filename setting from postgresql.conf.
Regards,
Andreas
Index: src/backend/postmaster/postmaster.c
==
Peter Eisentraut wrote:
Bruce Momjian wrote:
I have added this patch plus your later comments to the patch queue.
The autovacuum process still uses libpq to send its queries, which is
not the idea behind backend integration.
Can we consider this a non-fatal bug that has to be solved soo
Bruce Momjian wrote:
How is this patch supposed to work? Do people need to modify
postgresql.conf and then sighup the postmaster? It seems more logical
for the super-user to call a server-side function.
I assume calling pg_logfile_rotate() to be the standard way. calling
pg_logfile_rotate wil
Tom Lane wrote:
That was something that bothered me too. I think in the patch as given,
the GUC parameter determining the logfile name would have to be
PGC_POSTMASTER, ie, you could not change it on the fly because the
backends wouldn't all switch together.
In my original posting it was PGC_POSTM
Tom Lane wrote:
That struck me as not only useless but the deliberately hard way to do
it. To use that in the real world, you'd have to set up a cron job to
trigger the rotation,
Still on my radar...
which means a lot of infrastructure and privilege;
whereas ISTM the point of this feature was to
Bruce Momjian wrote:
Also there are no documenttion changes.
Here are the missing docs, freshly created against cvs.
Regards,
Andreas
Index: func.sgml
===
RCS file: /projects/cvsroot/pgsql-server/doc/src/sgml/func.sgml,v
retrieving rev
one is created (rotated).
* The logfiles are stored in a subdirectory (configurable in
* postgresql.conf), using an internal naming scheme that mangles
* creation time and current postmaster pid.
*
* Author: Andreas Pflug <[EMAIL PROTECTED]>
*
* Copyright (c) 2004,
CTION_ARGS);
+extern Datum pg_file_rename(PG_FUNCTION_ARGS);
+extern Datum pg_file_unlink(PG_FUNCTION_ARGS);
+
+extern Datum pg_dir_ls(PG_FUNCTION_ARGS);
+
/* not_in.c */
extern Datum int4notin(PG_FUNCTION_ARGS);
extern Datum oidnotin(PG_FUNCTION_ARGS);
Index: src/include/utils/elog.h
==
Magnus Hagander wrote:
> Super-minor nitpicking from just eyeing over the patch, not actually
> checking how it works.
Reviewing the own code the most obvious things are overlooked.
>
> This patch changes the error message for pg_signal_backend() to "only
> superuser may access generic file functio
Bruce Momjian wrote:
Andreas Pflug wrote:
Very nice. You did a nice trick of reading the log filenames into a
timestamp field:
count = sscanf(de->d_name, "%04d-%02d-%02d_%02d%02d%02d_%05d.log", &yea$
You only process files that match that pattern for pg_logfiles_ls()
(pe
Bruce Momjian wrote:
Peter Eisentraut wrote:
Bruce Momjian wrote:
Peter Eisentraut wrote:
Andreas Pflug wrote:
How should the prefix be named? pgsql_ ?
Make the file names configurable.
He has code to interpret the file names as timestamps that can be
used in queries. If we allowed full user
le configured
* in postgresql.conf. If these limits are reached or passed, the
* current logfile is closed and a new one is created (rotated).
* The logfiles are stored in a subdirectory (configurable in
* postgresql.conf), using an internal naming scheme that mangles
* creation time and c
Bruce Momjian wrote:
Andreas Pflug wrote:
OK, new idea. Forget about modifying pg_dir_ls(). Instead add
pg_file_stat the returns the file size, times. You can then easily use
that for file size and times. Also, if you want, add an is_dir boolean
so people can write functions that walk the
NCTION_ARGS);
extern Datum pg_tablespace_databases(PG_FUNCTION_ARGS);
+extern Datum pg_logfile_rotate(PG_FUNCTION_ARGS);
+extern Datum pg_logdir_ls(PG_FUNCTION_ARGS);
+
+extern Datum pg_file_stat(PG_FUNCTION_ARGS);
+extern Datum pg_file_read
Bruce Momjian wrote:
Here is what you can do:
SELECT filename,
(SELECT file_len FROM pg_file_stat(filename)),
(SELECT file_ctime FROM pg_file_stat(filename)),
(SELECT file_mtime FROM pg_file_stat(filename)),
(SELECT file_atime FROM pg_file_stat(filename))
FROM pg_dir_ls('/etc') AS
Bruce Momjian wrote:
Are we done? Seems pg_file_stat() works fine. Do we need other
adjustments?
Here are the documentation changes.
Regards,
Andreas
Index: catalogs.sgml
===
RCS file: /projects/cvsroot/pgsql-server/doc/src/sgml/ca
Tom Lane wrote:
Bruce Momjian wrote:
Are we done?
Nope, the syslogger part of this is still a mess. I don't want any
pg_logfile_rotate() function in there at all: its presence is a hangover
from a different design philosophy.
No. As I mentioned earlier, there might be cases when a superuser wants
Would you use a kill operation in the way you describe above
if you knew
that it had, say, a 1% chance of causing a database-wide PANIC
each time
you used it?
Seems there's the need for some connection killing functionality. If
it's not present, the whole cluster needs to be shut down, which mak
Tom Lane wrote:
If you don't mind plastering a "use at your own risk" sign on it, then
go for it.
killing a backend is obviously much more "at your own risk" than a
descent function.
Taken from your mail, I understand that a killed backend might leave
some loose ends, eg. open locks, which would
This is the known patch, with following changes:
- realStdErr handed over for EXEC_BACKEND, but still not tested
- Sometimes EMFILE is received in the logger's process queue, when a
backend ended after a SSL connection was interrupted. This is ignored
now (previously it forced an exit(1) and rest
gfile_rotate'
LANGUAGE C STABLE STRICT;
CREATE FUNCTION pg_logdir_ls() RETURNS setof record
AS 'MODULE_PATHNAME', 'pg_logdir_ls'
LANGUAGE C VOLATILE STRICT;
CREATE VIEW pg_logdir_ls AS
SELECT *
FROM pg_logdir_ls() AS A
(filetime
The attached patches will add
pg_get_ruledef(oid, int)
pg_get_viewdef(text, int)
pg_get_viewdef(oid, int)
pg_get_indexdef(oid, int)
pg_get_constraintdef(oid, int)
pg_get_expr(text, oid, int)
If the last parameter "pretty-print" is 0, these function will return
the same result as
Gevik Babakhani wrote:
> - new_guid() function is supported. This function is based on V4 random
> uuid value. It generated 16 random bytes with uuid 'variant' and
> 'version'. It is not guaranteed to produce unique values
Isn't guaranteed uniqueness the very attribute that's expected? AFAIK
ther
Dave Page wrote:
-Original Message-
From: Bruce Momjian [mailto:[EMAIL PROTECTED]
Sent: 01 August 2005 03:26
To: Dave Page
Cc: PostgreSQL-patches
Subject: Re: [HACKERS] For review: Server instrumentation patch
Dave Page wrote:
[Resent as the list seems to have rejected yesterday
Dave Page wrote:
pg_dir_ls isn't necessary for reading the logfiles;
pg_logdir_ls will do
this.
Err, yes, sorry - that was a thinko.
The list isn't complete. pgadmin uses these three functions for logfile
tracking:
- pg_logdir_ls to list logfiles
- pg_file_length to check for changes
Bruce Momjian wrote:
Dave Page wrote:
The only part I didn't like about the patch is the stat display:
test=> select pg_file_stat('postgresql.conf');
pg_file_stat
--
Bruce Momjian wrote:
I did not include pg_logdir_ls because that was basically pg_ls_dir with
logic to decode the log file name and convert it to a timestamp. That
seemed best done in the client.
IMHO omitting pg_logdir_ls is a bad idea, because the function is
intended to hide server inter
Bruce Momjian wrote:
BTW, it surprised me that one of the functions (don't remember
which one) expected the log files to be named in a very specific
fashion. So there's no flexibility for changing the log_prefix.
Probably it's not so bad, but strange anyway. Is this for
"security" reasons?
Tom Lane wrote:
Andreas Pflug <[EMAIL PROTECTED]> writes:
So how would your query to display all all available _logfiles_
look like?
I think it's perfectly reasonable to assume that all the files in the
log directory are logfiles --- more so than assuming that the admin
hasn'
Bruce Momjian wrote:
I don't assume people using psql will care about the current log files ---
Hm. Probably because you think these users will have direct file access?
Which in turn means they can edit *.conf directly too and don't need an
interface for that either.
it would be somethi
Bruce Momjian wrote:
I don't see how listing the log files relates to editing the confuration
files.
Both are remote administration. While we've seen the discussion that one
aspect (config file editing) should be performed in psql, you assume the
other aspect (viewing the logfile) to be no
Bruce Momjian wrote:
True, but that is more for the application. I don't imagine a user
looking at that from psql would have a problem.
However, you asked for a query that looks like pg_ls_logdir() and here
it is:
SELECT pg_ls_dir
FROM(
SEL
Tom Lane wrote:
Andreas Pflug <[EMAIL PROTECTED]> writes:
Bruce Momjian wrote:
Well, if they mix log files and non-log files in the same directory, we
would have to filter based on the log_filename directive in the
application, or use LIKE in a query.
.. which is what pg_logdir_l
Bruce Momjian wrote:
Also, do we have a way to return columns from a system-installed
function? I really don't like that pg_stat_file() to returns a record
rather than named columns. How do I even access the individual record
values?
As in pg_settings:
SELECT length, mtime FROM pg_file_stat
Tom Lane wrote:
I removed the separate pg_file_length() function, as it doesn't have any
significant notational advantage anymore; you can do
Please note that there are pg_file_length functions in use for 8.0 on
probably >95 % of win32 installations, so you're breaking backwards
compatibilit
1 - 100 of 178 matches
Mail list logo
